Author name cluster

Tarun Gupta 0002

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

4 papers

1 author row

IROS Conference 2023 Conference Paper

Hierarchical Imitation Learning for Stochastic Environments

Maximilian Igl
Punit Shah
Paul Mougin
Sirish Srinivasan
Tarun Gupta 0002
Brandyn White
Kyriacos Shiarlis
Shimon Whiteson

Many applications of imitation learning require the agent to generate the full distribution of behaviour observed in the training data. For example, to evaluate the safety of autonomous vehicles in simulation, accurate and diverse behaviour models of other road users are paramount. Existing methods that improve this distributional realism typically rely on hierarchical policies. These condition the policy on types such as goals or personas that give rise to multi-modal behaviour. However, such methods are often inappropriate for stochastic environments where the agent must also react to external factors: because agent types are inferred from the observed future trajectory during training, these environments require that the contributions of internal and external factors to the agent behaviour are disentangled and only internal factors, i. e. , those under the agent's control, are encoded in the type. Encoding future information about external factors leads to inappropriate agent reactions during testing, when the future is unknown and types must be drawn independently from the actual future. We formalize this challenge as distribution shift in the conditional distribution of agent types under environmental stochasticity. We propose Robust Type Conditioning (RTC), which eliminates this shift with adversarial training under randomly sampled types. Experiments on two domains, including the large-scale Waymo Open Motion Dataset, show improved distributional realism while maintaining or improving task performance compared to state-of-the-art baselines.

Details

ICLR Conference 2021 Conference Paper

RODE: Learning Roles to Decompose Multi-Agent Tasks

Tonghan Wang 0001
Tarun Gupta 0002
Anuj Mahajan
Bei Peng 0001
Shimon Whiteson
Chongjie Zhang

Role-based learning holds the promise of achieving scalable multi-agent learning by decomposing complex tasks using roles. However, it is largely unclear how to efficiently discover such a set of roles. To solve this problem, we propose to first decompose joint action spaces into restricted role action spaces by clustering actions according to their effects on the environment and other agents. Learning a role selector based on action effects makes role discovery much easier because it forms a bi-level learning hierarchy: the role selector searches in a smaller role space and at a lower temporal resolution, while role policies learn in significantly reduced primitive action-observation spaces. We further integrate information about action effects into the role policies to boost learning efficiency and policy generalization. By virtue of these advances, our method (1) outperforms the current state-of-the-art MARL algorithms on 9 of the 14 scenarios that comprise the challenging StarCraft II micromanagement benchmark and (2) achieves rapid transfer to new environments with three times the number of agents. Demonstrative videos can be viewed at https://sites.google.com/view/rode-marl.

Details

ICML Conference 2021 Conference Paper

UneVEn: Universal Value Exploration for Multi-Agent Reinforcement Learning

Tarun Gupta 0002
Anuj Mahajan
Bei Peng 0001
Wendelin Boehmer
Shimon Whiteson

VDN and QMIX are two popular value-based algorithms for cooperative MARL that learn a centralized action value function as a monotonic mixing of per-agent utilities. While this enables easy decentralization of the learned policy, the restricted joint action value function can prevent them from solving tasks that require significant coordination between agents at a given timestep. We show that this problem can be overcome by improving the joint exploration of all agents during training. Specifically, we propose a novel MARL approach called Universal Value Exploration (UneVEn) that learns a set of related tasks simultaneously with a linear decomposition of universal successor features. With the policies of already solved related tasks, the joint exploration process of all agents can be improved to help them achieve better coordination. Empirical results on a set of exploration games, challenging cooperative predator-prey tasks requiring significant coordination among agents, and StarCraft II micromanagement benchmarks show that UneVEn can solve tasks where other state-of-the-art MARL methods fail.

Details

ICAPS Conference 2020 Conference Paper

Reinforcement Learning for Zone Based Multiagent Pathfinding under Uncertainty

Jiajing Ling
Tarun Gupta 0002
Akshat Kumar

We address the problem of multiple agents finding their paths from respective sources to destination nodes in a graph (also called MAPF). Most existing approaches assume that all agents move at fixed speed, and that a single node accommodates only a single agent. Motivated by the emerging applications of autonomous vehicles such as drone traffic management, we present zone-based path finding (or ZBPF) where agents move among zones, and agents' movements require uncertain travel time. Furthermore, each zone can accommodate multiple agents (as per its capacity). We also develop a simulator for ZBPF which provides a clean interface from the simulation environment to learning algorithms. We develop a novel formulation of the ZBPF problem using difference-of-convex functions (DC) programming. The resulting approach can be used for policy learning using samples from the simulator. We also present a multiagent credit assignment scheme that helps our learning approach converge faster. Empirical results in a number of 2D and 3D instances show that our approach can effectively minimize congestion in zones, while ensuring agents reach their final destinations.

Details